Normative artificial intelligence for regulating manufacturing

• Grant number: 22/03454-1
• Support Opportunities: Regular Research Grants
• Start date: March 01, 2023
• End date: February 28, 2027
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Agreement: ANR
• Principal Investigator: Jaime Simão Sichman
• Grantee: Jaime Simão Sichman
• Principal researcher abroad: Luis Gustavo Nardin
• Institution abroad: École des Mines de Saint-Étienne, Gardanne (MINES Saint-Étienne), France
• Host Institution: Escola Politécnica (EP). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated researchers: Anarosa Alves Franco Brandão ; Eduardo de Senzi Zancul ; Fábio Levy Siqueira ; Gilberto Francisco Martha de Souza ; Jomi Fred Hubner
• Associated scholarship(s): 23/06395-9 - Modelling and Design for Regulating Manufacturing Systems - Phase I, BP.TT; 23/04322-4 - Prototypical deployments in manufacturing systems: phase I, BP.TT

Abstract

The digital transformation of manufacturing industries provides a nurturing environment for the adoption of more autonomous and (self-)adaptive technologies that can quickly and flexibly respond to endogenous and exogenous changes, while being transparent and complying with sustainable regulations. This modern manufacturing industrial setting is organised in three layers: (i) physical layer: access to the physical system, (ii) knowledge layer: information to manage and control the industrial processes, and (iii) application layer: an environment for automating these industrial processes. NAIMAN focuses on the last two layers, assuming that the heterogeneous physical systems on the physical layer are accessible via a uniform interface. In this complex ecosystem, industrial processes automation is tackled with the use of autonomous and intelligent agents that interact with each other on the application layer. In the knowledge layer, we target the domain knowledge representing the normative aspects (i.e., norms and sanctions) regulating these industrial settings and processes. Norms represent the expected agents' behaviour. We advocate that they are a rich and flexible concept for regulating manufacturing systems. Sanctions as reactions to any violation of or compliance with these expected behaviours are used to balance the agents' autonomy and the overall manufacturing system's control. The knowledge layer allows agents to reason about the production capabilities and capacities together with regulations to decide how and where to carry out their production tasks. The explicit normative representation and reasoning enable agents to both adapt the execution of industrial processes to unexpected situations and conditions, and to transparently and intelligibly express their decisions to an human operator. Hence, the main goal is to develop technologies demonstrated on industrial platforms that enable agents to operate in heterogeneous and dynamic industrial settings and reason about normative aspects to enhance flexibility, resilience, trustworthiness, and sustainability of manufacturing systems. (AU)